Service · bots
A turnkey Telegram bot
I write Telegram bots in Python (aiogram): from the dialogue flow to launch on a server.
your task
on Telegram
What goes into a turnkey bot?
- Flow: what the bot asks and what it answers
- Catalogue, cart and an invoice through Telegram Payments
- Scheduled digests and mailings
- Orders and requests go to a database, the owner gets a notice
- A check on test data and two rounds of changes
Which bots have I built?
- “What’s up?” — every morning the bot sends what people search for in your field, and why.
- Shop bot — catalogue, cart and Telegram Payments without leaving the chat.
How the work goes
- 01ConversationWe discuss the task. The spec and the list of API connections grow out of it.
- 02StructureWhat goes where and why. Agreed before design.
- 03DesignMockups of the key screens, to understand and estimate.
- 04MVPA working version at a test address: click through it, show it to colleagues.
- 05ConnectionsPayments, CRM, booking and bots are checked on test data.
- 06PolishMetrica and two rounds of changes. Bugs and typos are fixed free, with no time limit.
Questions
- Need a spec?
- No, we write it together.
- How much?
- After a call: plan, timeline and price.
- Contract?
- Yes, I work under contract.
- Prepayment?
- Half before the start, half on delivery.
- What if it is urgent?
- I take it. Name the deadline and I will say if I can make it.
- Whose access is it?
- Yours. Domain, builder, Metrica and mail live in a separate mailbox that I hand over with the project. Grow it with me, run it yourself or pass it to another team.
- What is paid separately?
- The domain and the builder or server plan — directly, so the asset is yours from day one. I name the rough yearly sum before we start.
- Changes after launch?
- Yes, free of charge: I refine by your comments. New sections and features are priced separately.
Describe your task in Telegram — after we talk I name the plan, the timeline and the price. Or by e-mail: vx@marcellapng.com
Related — in plain words
- How a neural network works — how it learns from examples, what weights and tokens are, why it gets things wrong.